2000 hours in experiment 2. The fish might have been disturbed at this time because of frequent 

 switching on and off of lights in adjacent rooms. The pattern for experiment 1 deviates from that 

 of the other two experiments, possibly because of greater fluctuations in the temperature of the 

 water (table 3), The results of this series, at any rate, are to be interpreted cautiously. 



Influence of Tennperature on Rate of 

 Oxygen Consumption 



Procedure. --With the arrange- 



ment shown in figure 2b, it was possible to 

 raise the temperature in the contadner by ap- 

 proximately 2 intervals and keep it practi- 

 cally constant at aijy one level. The average 



variation in temperature was 0. 3 C, , the 



o 

 maximum variation in any test being 0. 5 C, 



All experiments were begun at 19 C. , at 

 least 2 hours being taken to lower the tem- 

 perature to this value. One set of Seimples 

 was withdrawn at each temperature level, 

 not less than 1/2 hour after the temperature 

 reached the desired value. Fish used ranged 

 in length from 34-50 mm. , the average 

 length of fish in each experiment ranging 

 from 42-48 mm. The tests were run over a 

 -day period in October 1951, during which time the temperature in the stock tanks was about 



'c. 



1800 2400 



TIME OF DAY 



Figure 3. 



-Rate of oxygen consumption 

 (cc. /gmVhr. ) in iao over a 

 24-hour period. 



Results. --The curve of temperature versus rate of oxygen consumption (fig. 4) 

 resembles those found for other fish (e.g., Ege and Krogh 1914; Wells 1935a; Haugaard and Irving 

 1943). It rises slowly at first, the slope of the curve being somewhat steeper at the higher tem- 

 peratures. Over the range in temperatures tested (19 -29 C. ) there is an approximately twofold 

 increase in the rate of oxygen consumption. Wells (1935a) has pointed out that small Fundulua 



show a more pronounced response to tem- 

 perature changes than larger fish; how- 

 ever, the occasional presence of smaller 

 iao (34-37 nnm. ) in the present experi- 

 ments did not appear to alter the pattern 

 of response to temperature. Knowing that 

 the previous temperature history of the 

 fish (acclimatization temperature) can 

 markedly affect subsequent response to 

 temperature changes (Wells 1935b, c; 

 Sumner and Doudoroff 1938; Sumner and 

 Lanham 1942), this factor was controlled 

 as far as possible by running the experi- 

 ments during one month (October) on fish 

 which had been kept for some time at 

 about 26 C. 



Influence of Flow Rate and Numbers 

 of Fish upon the Metabolic Rate 



Procedure . --In a typical 

 '^ ^' ^' ^^ ^' ^^ *' experiment, the approximate desired num- 



TEMPERATURE °C ^^^ °^ fish were transferred to the experi- 



mental jar (fig. 2c), the flow was adjusted 

 Figure 4. -- Rate of oxygen consumption to an intermediate value, and about 1 hour 



(cc. /gm. /hr. ) at different was allowed for the fish to "settle down" 



temperatures. 



